Over the years of research, EEG signal study has grown to give promising outcomes. A lot of research has been done on implementing brain-computer interfaces, detecting seizure through EEG signal abnormalities and so on. A stimulus is given to trigger the EEG signals, the resulting responses are recorded, analyzed and inferences are drawn. Brain signal analysis is an important aspect in understanding signal properties by extracting useful information that describes the signal. Using these extracted features, the EEG signals can be categorized based on the patterns, abnormalities or uniqueness that they may reflect. EEG signal classification finds great use in clinical applications and prosthetics. This study aims at delivering an efficient algorithm to classify EEG signals using signal analysis methods and machine learning techniques. An EEG signal is first simulated and tested with multiple scenarios to generate different types of patterns in them. Signal preprocessing and feature extraction techniques are applied to the signal. Specifically, the strength of the signal is computed using root mean square (RMS) methods and classified using neural network models. Accordingly, a new index called R-index is introduced to increase classification accuracy. This series of processes is then tested upon human data to validate the robustness of the proposed algorithm. EEG signals are collected using state-of-the-art signal acquisition system, g.Nautilus. In this study, voiced sounds have been used as auditory stimuli. Various pitches and patterns have been incorporated in the stimuli to generate different EEG patterns for analysis. Temporal and prefrontal regions of the brain have been targeted, thereby using FP1, FP2, T7 and T8 channels of the EEG signal acquisition system. These are analyzed in a fashion, similar to the simulated signals. The acquired data was analyzed using RMS analysis methods and hierarchical clustering analysis (HCA) techniques. Statistical significance of the data was strengthened using Analysis of Variance (ANOVA) approach. In a nutshell, the objective of this study was to formulate an efficient algorithm using different techniques to identify different patterns in EEG signals using artificial neural networks.